Article Text

Download PDFPDF

The value of microarray-based comparative genomic hybridisation (aCGH) testing in the paediatric clinic
  1. T A Briggs1,
  2. J Harris1,
  3. J Innes1,
  4. A Will2,
  5. P D Arkwright3,
  6. J Clayton-Smith1
  1. 1Manchester Centre for Genomic Medicine, St Mary's Hospital, University of Manchester, Manchester, UK
  2. 2Department of Paediatric Haematology, Royal Manchester Children's Hospital, Manchester, UK
  3. 3Department of Paediatric Allergy & Immunology, University of Manchester, Royal Manchester Children's Hospital, Manchester, UK
  1. Correspondence to Dr T A Briggs, Manchester Centre for Genomic Medicine, St Mary's Hospital, University of Manchester, Manchester M13 9WL, UK; tracy.briggs{at}

Statistics from


Many human genetic disorders result from an unbalanced chromosomal abnormality in which there is a loss or gain of chromosomal material. Since the 1970s, such imbalances have traditionally been detected through the analysis of chromosomes by karyotype assessment. Chromosome G-banding patterns are visualised under the microscope leading to the identification of differences in chromosome number (aneuploidies), large balanced and unbalanced structural rearrangements and mosaic structural and numerical abnormalities. Karyotyping has been an integral tool in the genetic evaluation of children with congenital anomalies and developmental delay. However, chromosomal aberrations below approximately 5 megabases (Mb) in size cannot be detected by routine karyotype and therefore, a loss or gain of genetic material below this size (microdeletion or microduplication) cannot be visualised.

The diagnosis of chromosome abnormalities below 5 Mb can be improved using fluorescently labelled DNA probes, which bind to specific DNA sequences. This technique, called fluorescence in situ hybridisation (FISH), is valuable in the identification of chromosomal aberrations at predetermined locations. For example, the absence of a fluorescent probe signal on one copy of chromosome 22 at the q11.2 locus indicates 22q11.2 deletion syndrome (see table 1).1 While FISH is very useful, it is a targeted test that requires a specific diagnosis and chromosomal location to be considered. A more hypothesis-free approach is achieved by using microarray-based comparative genomic hybridisation (aCGH). A description of how aCGH is performed was recently published.2 In brief, thousands of probes spread across all chromosomes are used to interrogate chromosome copy number and any deviation from the expected number of copies is reported, such that a locus in which a group of probes are missing indicates a deletion of material, while extra copies are consistent with a gain of material. These variants are termed copy number variants (CNVs) collectively and may represent normal human variation or may be associated with disease.

Table 1

Chromosomal deletion and duplication syndromes

A gain of genetic material may be less likely to cause disease than a loss; however, one factor that can influence the possible effect of a duplication is the location of the additional material. The duplicated material may be adjacent to the original copy on the chromosome (ie, in tandem), or may have inserted into a different location on the same or a different chromosome, or even be present as a supernumerary chromosome (termed a marker chromosome). This positional information cannot be determined by aCGH, but can be elucidated with using a FISH probe specific for the duplicated region. Position is important because relocation into another site may interrupt gene function at that locus and cause a phenotype additional to that, which may be associated with increased gene dosage.

The value of aCGH was evaluated by the International Standard Cytogenomic Array Consortium,3 which found that in children with unexplained developmental delay/intellectual disability, autistic spectrum disorders or multiple congenital anomalies, a causative CNV was detected in 10%–15% of patients by aCGH, compared with only 3% by karyotype analysis (excluding Down syndrome and other recognisable chromosomal syndromes). This improved diagnostic yield has an obvious positive impact on both patient care and economics and aCGH is therefore the recommended first-tier test for individuals with developmental disability or congenital anomalies.3

As larger numbers of aCGH studies are undertaken, areas of recurrent CNVs, associated with specific phenotypes are emerging. Following the increasing recognition of these recurrent microdeletion and microduplication syndromes, online patient information leaflets have been produced by UNIQUE, a charity founded for families with rare chromosome disorders. These leaflets describe both the basis of microarray technology and clinical summaries of a number of recurrent CNV syndromes ( Table 1 describes some of the more frequently detected CNVs and where known, lists the genes within these regions which are considered the key causative genes in the manifesting phenotype. The phenotypical descriptions in the table highlight the fact that the features of many CNV syndromes overlap and a specific locus may not be clinically apparent to the referring clinician, thus illustrating the diagnostic utility of aCGH over FISH.

Recent studies have shown that complex genetic mechanisms, with combined chromosomal and single gene variants may cause genetic syndromes. An example is thrombocytopenia absent radius syndrome, in which both a microdeletion at the 1q21.1 locus on one allele (detected by aCGH) and a low-frequency single nucleotide polymorphism within a gene regulator on the other allele (detected by sequencing) are required to cause the disorder.15

Array CGH is of great value in paediatrics in the study of developmental disabilities and multiple congenital anomalies,2 but it increasingly recognised that the technique may also have wider relevance to general paediatric disorders. To illustrate this, the clinical case described below highlights the value of microarray in the immunology and haematology clinic.

Case report

A Caucasian boy was the second child of unrelated parents. He was born at term, weighing 3.4 kg (25th–50th centile). He was 3 months old, when he presented with poor weight gain (2nd centile) and a macrocytic anaemia, (haemoglobin 44 g/L; normal 105–135); mean corpusclar volume 86 fL (normal 64–84), but normal leucocyte and platelet counts. A blood film showed teardrop polikocytes, spherocytes with a poor reticulocyte response of 7×109/L (normal 8–105). B12 and folate were normal and the ferritin level was slightly elevated at 282 μG/L (normal 15–200). A chest X-ray showed cardiomegaly. Subsequent investigations excluded hereditary spherocytosis, pyruvate kinase and G6PD deficiency, haemoglobinopathy and parvovirus B19 infection. A bone marrow aspirate and cytogenetics were also normal.

Haemoglobin concentration remained low at 79–92 g/L (normal 101–138) and and at 22 months of age the patient was reviewed in the immunology clinic, following a one-month history of a chest infection. His weight was below the 0.4th centile, while his height and head circumference were on the 3rd centile. He had rather unusual, straight thumbs, but was not facially dysmorphic, and though he only had a few intelligible words, there were no other concerns regarding his development. In addition to the persistent mild anaemia, repeat blood count demonstrated a low white cell count at 2.6×109/L (normal 6–17), with a neutropenia of 0.5×109/L (normal 1.0–6.0). Epstein–Barr virus serology was consistent with active infection. He was given co-trimoxazole in view of the neutropenia and chest symptoms. Due to the cytopenia, failure to thrive and speech delay, Mitomycin C studies and aCGH were requested. There was no evidence of chromosomal fragility following Mitomycin C studies and thus a diagnosis of Fanconi anaemia was considered highly unlikely. However, aCGH demonstrated an approximately 2 Mb deletion on the long arm of chromosome 3 at the q29 locus (3q29).

Implications of the aCGH result

3q29 microdeletion syndrome is a recognised entity, with the microdeletion typically being around 1.6 Mb in size and associated with a variable phenotype. Affected individuals are usually healthy, without major birth defects, although kidney abnormalities and cleft lip and palate have been reported. There is an increased incidence of speech delay, as observed in our case. Gross motor delay, mild to moderate mental retardation and autism are reported in some individuals, as is an increased prevalence of childhood and adult psychiatric diagnoses.16 Some patients have subtle dysmorphic features (long face and nose, short philtrum and large ears), and 50%–60% of children are microcephalic.6 ,17

Assessment of the genes within a CNV may help to understand the phenotype. The recurrent 3q29 microdeletion region contains 21 genes, including PAK2 and DLG1, which are autosomal homologues of two X-linked mental retardation genes, PAK3 and DLG3. PAK2 and DLG1 are thus postulated to be causative of the mental retardation.6 These two genes and a third, FBXO45, have also been hypothesised to be causative of the psychiatric manifestations, as they have a role in synaptic transmission.16

In the majority of cases, 3q29 microdeletions arise as new mutations with a low recurrence risk. In a minority, however, the deletion is inherited from a parent, either directly or as an unbalanced form of a balanced rearrangement, in which case the risk of recurrence is high. In view of the variability of the condition, normal and very mildly affected parents who have not come to medical attention are described17 and therefore parental testing is indicated in all cases. Parental testing in our case demonstrated that the deletion was de novo.

Anaemia is not usually associated with 3q29 microdeletion syndrome, but the deletion in this child was larger than usual and contained six additional genes, including the RPL35A gene. The ribosomal protein L35A encoded by RPL35A has a role in RNA processing, ribosomal biogenesis, selective cell proliferation and apoptosis. Deletions and mutations in RPL35A are reported as causative of Diamond–Blackfan anaemia (DBA)18 and deletion of this gene therefore explained the patient's haematological phenotype. DBA is a bone marrow failure syndrome, associated with anaemia and occasionally neutropenia and thrombocytopenia. Additional features of short stature and congenital anomalies are reported in 40% of cases, including craniofacial, cardiac, genito-urinary and upper limb malformations. There is also an increased incidence of haematological malignancy and osteogenic sarcoma. This is therefore an important diagnosis to make in terms of further surveillance and management.

DBA is caused by mutations in at least 10 different genes encoding ribosomal proteins18–20 and usually follows an autosomal dominant pattern of inheritance. Landowski et al21 conducted aCGH in 87 probands with DBA who did not harbour point mutations in 10 known DBA-associated genes and identified deletions in 7% of cases involving five different genes, suggesting that aCGH may be of diagnostic value. Indeed, the role of RPL35A in DBA was first identified following an interrogation of the deleted genes in two patients with DBA who had 3q29 deletions.18 Furthermore, we are aware of a second case of RPL35A related DBA22 where, as in our case, the diagnosis was not made until aCGH was performed. This apparent diagnostic delay may be explained in our case by the fact that the bone-marrow was not diagnostic, perhaps indicating an attenuated disease form. More RPL35A deletion cases are needed however before any genotype–phenotype correlations can be made.

This case highlights the clinical utility of aCGH in children with unexplained anaemia, particularly if there are additional features. The case demonstrates the clinical variability of known syndromes associated with CNVs and how the phenotype may be more diverse if the copy number change identified is larger than usual. The importance of parental testing to allow provision of appropriate family genetic counselling should also be noted, particularly as phenotypical severity often varies and parents may not have features of the syndrome.


As demonstrated above and in table 1, aCGH can be of great value in a diverse number of situations, both in making a diagnosis and then, with an understanding of reported risks associated with a syndrome, appropriately managing the child and family. It is important to note, however, that analysing and interpreting aCGH findings can be complicated and advice from the cytogenetics team and referral to a clinical geneticist may be required, a fact also acknowledged by Kharbanda et al.2

Prior to initiating aCGH, it is important to ensure that discussion as to the possible outcomes of the test are undertaken. These are:

  1. No currently plausible pathogenic variant identified;

  2. Pathogenic or likely pathogenic variant identified, which may rarely be an incidental finding;

  3. Variant of unknown significance identified.

In our laboratory over the last 7 years, approximately 10% of cases tested with aCGH have been reported to have a ‘pathogenic or likely pathogenic variant identified’. This CNV may be a recurrent microdeletion/duplication syndrome and information about the specific associated syndrome should be available. As in the above case illustration, further assessments are warranted for the parents/extended family where possible. When parental testing is undertaken in our laboratory, approximately 50% of the recurrent CNV syndromes identified have been inherited. In some cases this may explain symptoms in parents, while other parents may be asymptomatic, illustrating the highly variable nature of these syndromes and the difficulties providing counselling within the family, even for recurrent syndromes.

In many cases CNVs detected by aCGH are not in well-described, recurrent regions and there may be no published literature describing similar findings. In these cases, it can be extremely difficult to know if such changes are the cause of the phenotype, or not. However, CNVs are likely to be reported as ‘pathogenic or likely pathogenic’, if they are large, or if there are multiple losses or gains of genetic material, or if the CNV encompass a known gene where abnormal gene dosage is known to cause disease, such as in our case of DBA. Parental studies are again recommended in these cases to inform genetic counselling.

Finally, as indicated above, a ‘pathogenic or likely pathogenic variant’ may be an incidental finding, in that the CNV is not related to the presenting symptom, but may infer a current or future health risk, and so is of clinical relevance. For example, the finding of a deletion of a known adult onset cancer predisposition gene will be of relevance to the child in the future and if inherited, will have more immediate implications for the carrier parent. The frequency with which such findings are identified is low, particularly in the absence of a family history of other diagnoses, but this possibility should be raised prior to testing. If an incidental finding is discovered, further testing in the family is indicated and very likely a referral to Clinical Genetics.

In some cases, a ‘variant of unknown significance’ may be identified and by definition the pathogenicity of such variants cannot be determined based on current knowledge. In terms of clinical management, in such cases, a paediatrician should consider alternative diagnoses, which may include a referral to Clinical Genetics. Over time, the significance (or lack) of specific chromosomal gains or losses should become clearer, particularly due to the development of CNV databases, which have been established to record both cytogenetic data and, additionally in some cases, fully anonymised phenotypical information (with parental or patient consent) (see Kharbanda et al2).

Truly balanced chromosomal rearrangements and low-level mosaicism are generally not detectable by aCGH, but these are relatively infrequent causes of abnormal phenotypes in the paediatric population (<1%). There are, however, certain cases in which aCGH may not be the most suitable first-line tool namely:

  1. In a child who has an obvious aneuploidy, such as Down's syndrome, rapid QF-PCR or karyotype analysis should be requested.

  2. If there is a family history of specific genetic disorder targeted single gene or chromosomal evaluation is recommended.

  3. In cases of ambiguous genitalia, QF-PCR followed by a karyotype are required in initial management to determine the genetic gender and to look for apparently balanced structural rearrangements.

While acknowledging the potential counselling and analytical difficulties associated with aCGH, (beyond the three scenarios outlined above), we would recommend that aCGH should be used as the first-line test in:

  1. Children with developmental delay/intellectual disability or a congenital anomaly (including if the diagnosis remains unknown in those who have previously had karyotype analysis).

  2. Children (or a family) with an unusual combination of clinical features, in which it is difficult to provide a unifying diagnosis—such as in the case of DBA we described above.

Identification of chromosomal diagnoses by karyotype and now aCGH is of great value to families in terms of improving diagnosis and management for their child and to facilitate improved understanding of recurrence risks in future pregnancies. Furthermore, aCGH has allowed the delineation of new syndromes, and the study of genes within the deleted/duplicated regions has enhanced our understanding of gene function.


We would like to acknowledge the family of our patient for their support in writing this manuscript.



  • Contributors TAB managed the case from a clinical genetics perspective and wrote the paper. JI and JH performed the genetic studies and assisted with writing the paper. JCS edited the paper and provided genetic advice. PA and AW medically managed the case described and AW and PA edited the case description.

  • Funding Dr TAB is funded by a Clinical Lectureship from the National Institute for Health Research.

  • Competing interests None.

  • Patient consent Obtained.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.